Ecommerce SEO

CHAPTER 3

Keyword Research

Length: 10,599 words

Estimated reading time: 1 hour, 10 minutes

This e-commerce SEO guide has almost 400 pages of advanced, actionable insights into on-page SEO for ecommerce. This is the third out of 8 chapters.

Written by an e-commerce SEO consultant with over 25 years of research and practical experience, this comprehensive SEO resource will teach you how to identify and address all SEO issues specific to e-commerce websites in one place.

The strategies and tactics described in this guide have been successfully implemented on top 10 online retailers, small & medium businesses, and mom-and-pop stores.

Please share and link to this guide if you like it.

Keyword Research

While SEO is the abbreviation for “search engine optimization”, SEO experts do not improve how search engines work; they optimize for search engines. And because the primary purpose of search engines is to be helpful to the people who use them, SEO would be better thought of as optimizing your website for search engine users. Search engine users are also referred to as searchers.

The search trifecta includes three entities:

  • The user
  • The search engine
  • The website

A common search experience looks like this: The user enters a search query on the search engine, which leads searchers to the website, which should (in an ideal state) fulfill the user’s query.

Figure 54 – The search trifecta.

We often skip the user and jump straight to the search engine when performing keyword research. This section describes what I believe is a better approach to keyword research: start with the user, continue with the website, and finally, consider the search engine.

I will refer to keywords and queries interchangeably, but there is a subtle difference between them.[1] A search query is a series of words users type into a search engine. A keyword is an abstract concept within a search query.

Figure 55 – Keywords are abstract concepts.

For example, on e-commerce websites, keywords are represented by department, category, or subcategory names. A search term that contains several words, including the keyword, is a search query.

Good information architecture and keyword research are at the foundation of great ecommerce websites that perform the best in search engines and convert at high rates. In the Information Architecture section, we found that deciding on primary and secondary navigation labels (or category and subcategory labels) based solely on keyword research is not optimal—it should be complemented by user testing and research or by using controlled and custom vocabularies. That is because the user intent is not always reflected in what they type in a search engine. That is also why estimating user intent by analyzing keywords or search queries isn’t easy.

Keyword or query research is a core concept for e-commerce websites because it is important for both users and search engines to map keywords with the right type of content. Discussing search engines and keywords outside the context of users is not the correct SEO approach.

In marketing, research means collecting all the raw data you will later use to perform an analysis. About keywords, research means collecting keyword data from different sources. Here are some data and metrics that you might collect:

  • The keywords or the search queries used by searchers on search engines.
  • Their associated search volumes.
  • The current rankings (keep in mind that rankings are difficult to measure accurately due to personalization and geo-location).
  • Competitiveness data such as the average DomainAuthority of the top 10 ranking domains.

You will collect the above keyword data directly from search engines or by using third parties.

Gathering keywords

Collecting the initial set of keywords is straightforward, but the number of potential sources is overwhelming. You can use the following:

  • Google’s Keyword Planner.
  • Google’s Display Planner.
  • Google’s autosuggest feature (crank it up with Ubersuggest or Keyword Snatcher).
  • Google and Bing’s related searches.
  • Bing’s Keyword Research feature within Bing Webmaster.
  • You can collect keywords by brainstorming with various internal departments or using your existing Google Ads campaigns.
  • You can also use Google Search Console data.
  • Social media sources (Twitter, Facebook, LinkedIn, etc.).

Other less-known sources for collecting keywords are:

  • Internal site search data using Google Analytics or other web analysis tools.
  • Voice-of-the-customer surveys and research.
  • User testing.
  • The anchor text of the natural links to your pages.
  • Competitor analysis.

Even though there are many keyword tools, the most extensive set of keywords and search queries and the most accurate search volumes can be extracted from pay-per-click (PPC) advertising platforms such as Google Ads.

I recommend collecting keyword data using an active Google Ads campaign rather than the data Google Ads provides without a live campaign. This is because when you run a live campaign, the Google Ads data goes beyond the keyword suggestions within the Keyword Planner. A live campaign can generate a handy list of long-tail keywords (use the Search Terms Report in Google Ads), and in my experience, that list is impossible to capture with any other tool.

Besides Text Ads, you should run Product Listing Ads in Google Ads (via the Merchant Center) and Dynamic Search Ads and then use the Search Query Report to get a fantastic number of relevant keywords. Many of those keywords will be long-tail keywords.

Figure 56 – It is easier to rank for a search query with more words (long-tail keyword) because the search query is usually less competitive.

Unfortunately, many marketers stop keyword research after collecting only quantitative data, such as search volumes. I call this the traditional keyword research approach.
In the digital marketing world, the following is a typical scenario:

“We identified that these keywords have the highest search volumes, so we should target them. We will change page titles, we will go with a 3% keyword density, and we will build a bunch of backlinks to the pages targeting those keywords”.

Alternatively, if the marketing person or agency is more knowledgeable, the scenario may sound like this:

“These keywords have a decent amount of traffic and have good conversion rates, as per your analytics data. They are competitive, and that is why we should optimize the internal linking and build backlinks to the most appropriate SEO’d pages”.

Yes, search volume data research is necessary, but you need to go much deeper than this if you want to increase organic traffic. Search volumes are just the starting point. Considering your users’ concerns, questions, and FUDs (fears, uncertainties, and doubts) would be best. All of these affect their purchasing decision. Once you identified those pain points, create content that attracts qualified organic traffic and generates sales.

Seasoned marketers call this concept Intent to Content.

Creating Personas

One of the best ways to map “intent-to-content” is by creating personas. A definition of persona is “a quasi-imaginary representation of your customer based on market research and real data about your existing customers. A persona includes demographics, behaviors, motivations, and goals”.

Ecommerce businesses, especially in the B2B space, need to go above and beyond and develop well-researched buyer personas to attract people in the early stages of the buying funnel. You must create and market content for every stage of the buying funnel.

Let’s say that you sell promotional products to businesses. Here’s what an oversimplified persona creation could be like:

Start by identifying the segments you need to market to and by giving them names, for example:

  • Vera, the Marketer.
  • Chris, the IT Geek.
  • Brad, the Economic Buyer.

You will focus on Vera, the Marketer, if you sell promotional products.

Creating Vera’s buyer profile should be comprehensive. Everyone involved with marketing and sales should contribute to a Persona Questionnaire based on their experience, knowledge, and online research of Vera. You can also interview existing customers who share Vera’s profile to find commonalities. A joint marketing and sales team should develop this questionnaire and should include questions such as:

  • Where does Vera read online?
  • Where does she go to ask for help?
  • What kind of wording does she use online?
  • What challenges does Vera face right now?
  • What are her goals?
  • What does her career path look like?
  • What motivates her to select a competitor?
  • How does she make decisions?

Additionally, you can collect and analyze public résumés to identify career paths for people involved in marketing decisions, like Vera. Here’s what the word cloud for marketing managers’ accountabilities may look like:

Figure 57 – Responsibilities for marketing managers.

Some of the most important facts you need to uncover about Vera are her pain points and how she makes decisions. You will reveal such data by engaging on websites where she goes to read, educate herself, or ask for help.

Once you have identified Vera’s pain points and top challenges related to your vertical (in our example, promotional products), bucket them into different content types and rank them based on the most severe problems. Then, prepare content to address each pain point or challenge (e.g., case studies, how-to’s, extensive guides, etc.).

One type of content for the upper buying funnel can be targeted to raise awareness about a given challenge. For the mid-funnel, the content might be a Guide on addressing the same challenge, with examples of how you helped other businesses deal with the problem. The content can be a case study for those ready to make a vendor selection. However, none of these should be salesy, just excellent and useful information.

You will identify Vera’s most important problems, educate her, and prove you have the products she needs.

Note: usually, it is a bad idea to “gate” upper and mid-funnel content, for example, by asking for email and contact details to access the content.

The Intent-to-Content concept became more relevant and prominent after the Hummingbird algorithm update. This update focused on processing conversational queries, which are longer, question-like queries including modifiers such as how to, where is, or where can. The focus shifted away from the traditional word-parsing approach.

Another objective of Hummingbird was to match the user intent so that Google could provide answers rather than just search results. The intent to content concept became even more important after introducing the so-called “position zero”, also known as the quick answer box.

We need to discuss the buying funnel and the user intent to map keywords-to-intent and intent-to-content.

The buying funnel

In the US, e-commerce conversion rates are about 3%[2] because online retailers focus mostly on converting branded traffic and marketing to consumers in the late stages of the buying funnel. Also, ecommerce websites usually concentrate their link-building campaigns on category and subcategory anchor text.

Keyword research and web analytics tools, PPC data, and other similar sources provide insights on which keywords the target audience searches for, when, and where they search from. However, savvy online retailers must understand the searcher context and create a content plan accordingly. They try to answer the why question. User testing is one great way to gauge the why, but it has limits.

Users do not just turn their computers on, type in your website name, and buy products from you. First, they realize they have a need; next, they research online,[3] decide what’s right for them, and—only then—purchase.

This journey is called the buying funnel.

Figure 58 – The stages of the buying funnel.

In the image above, you can notice how keywords in the awareness stage are generic and broad. They gradually become more specific until they finally become the product the searcher wants to purchase.

Although the keyword categorization in this example may seem straightforward and logical, in practice, the keywords used by consumers will belong to multiple categories and be found in various buying stages. That is why you need not be too particular about where to bucket a keyword.

Here are some great insights from one study that mapped 40,000 PPC keywords to the buying funnel:[4]

  • Targeting only keywords in the Purchase and Decision stages of the buying funnel for ecommerce websites can, theoretically, lead to 79% less organic traffic.
  • The buying funnel is representative of actual online consumer behavior, at least at the individual query level (6. Discussion and Implications, p. 11).
  • Advertisers can use the model to organize separate campaigns targeting various consumer touch points (6.3. Practical Implications, p. 14).
  • The implication is clear: do not ignore Awareness key phrases (6.3. Practical Implications, p. 14).

The researchers analyzed data from about seven million keywords from a large retail chain having a brick-and-mortar and online presence:

Figure 59 – The stages of the buying funnel[5].

As you can see in this screenshot from the study, awareness and research keywords comprise almost 80% of the total keywords. The same research indicates high PPC advertising costs for the Awareness (25%) and Research (57%) stages. A staggering $4.6 million (57% of $8 million) could have been saved with a proper keyword-to-content strategy.

When you map keywords to buying stages and to user intent, and when you develop content accordingly, you will:

  • Generate content that attracts organic traffic.
  • Create content that can be linked to more easily.
  • Support pages in the vertical silos up to top-level categories.
  • Reduce advertising costs.

The awareness stage

This is the first stage in the buying funnel. Your customers realize that they have a need or a problem, and they start researching general information about what would help them fill that need or fix it. They want to know what products or services are available on the market.

For an ecommerce website, the queries that can be associated with the awareness stage are the broadest, most generic terms, such as department, category, or subcategory names (e.g., “commuter bikes”, “winter jackets”, “car racks”, “running shoes”, “cruise deals”, “diamonds”, and so on).

However, longer and natural language queries are also found at the awareness stage. For example, a searcher wants to know how to save on his daily commute. He starts by typing “best ways to save on commuting”, then reads an article about commuter bikes.

At this stage, consumers do not know yet what will address their needs and are still seeking information, so awareness queries usually contain neither brand nor specific full product names. They can include an action or a problem that needs to be solved—e.g., “removing wine stains”.

According to the same study cited earlier,[6] an awareness search query:

  • Does not contain a brand name.
  • Could contain the partial product name/type.
  • Could contain the problem to be solved.

At this stage, the user intent is mostly informational.

Tactics for this stage
The search queries associated with the awareness stage are what most of the e-commerce websites target, for example, head terms such as department, category, subcategory, or sub-subcategory names. These search terms are usually super-competitive, and realistically, ranking for such keywords will not happen unless your website has a significant amount of authority and reputation in the industry (including backlinks). You will also need a lot of great content that establishes your website theme and makes you the go-to resource for a subject matter or theme (i.e., wines).

Some tactics associated with attracting traffic for these keywords are:

  • Creating content such as community pages, how-to pieces, blog articles, educational content, and linking vertically from such content. You will need significant content and consistent linking to support category pages.
  • Siloing the website with directories and internal linking.
  • Building themed backlinks to the content-rich pages and articles.
  • Showcasing instructional videos by featuring them on your website and through social media.

The research stage

At this point, the consumer has identified the type of product or service that could help. The possible customer can recognize brands in your industry or niche but has not yet decided on a definitive brand. The customer still needs to refine their knowledge before making a purchasing decision.

While the search queries are still broad, the consumer uses more specific terms, including keyword modifiers such as brand names or geo-locations, instead of generic searches.

The queries may look like “lightweight commuter bikes”, “insulated winter jackets”, “rear mounted car racks”, “cross training running shoes”, “European cruise deals”, or “4-carat wedding ring”. The queries can be subcategories, sub-subcategories, or product attributes.

Long-tail queries can be found at this stage as well. In our biking example, the consumer may type: “What are the best brands for commuter bikes”, “Which brand is more reliable”, “Compare {brand1}with{brand2}”, “what bike size do I need”, “what is the cost of an electric bike”, “how much will it cost to maintain a bike”, and so on.

At this stage, the user intent is still mostly informational, but transactional intent may be there, too.

Tactics for this stage

  • Write product reviews, product comparisons, and many articles to answer your target market’s questions.
  • Write extensive user guides (e.g., “How to Select an Electric Bike” or “How to Choose a Commuter Bike in 10 Easy Steps”). This type of content is an organic traffic driver and can potentially become a real link magnet. Then, promote this content socially and with influencers in your industry to generate buzz and, hopefully, backlinks. If you are a small or medium business focusing on a line of products or niche, there is some encouraging news for you. Because the authoritative websites you compete with on Google have a lot of inventory and themes to create content for, you may have an advantage if you are focused.
  • Create buyer personas to identify a) where the target market goes to read information and b) what questions they have. Once identified, group them into topics and write articles to answer the questions. To learn more about personas, read this leaked document about BestBuy’s buyer personas.[7]
  • Keyword-rich internal linking is also crucial at this stage. Internal linking is essential at all buying funnel stages, so make sure you cross-link from informational pages to category and product details pages.

The decision stage

Now, your prospective customer knows what solution is good for them. The prospect will research the best store to buy from, and it will try to get the most value. His logic and emotion will favor a particular brand. The possible customer will be much closer to making a purchase decision.

At this stage, the consumer has chosen a product and a brand but not the exact model number or version of the product. In our bicycle example, the searcher wants the Ridgeback brand and needs a commuter bike.

The Decision stage is where comparison shopping occurs, so search queries often include brand names and technical specifications. At this point, his queries will be more focused than in the previous two stages and can consist of very strong commercial intent keyword modifiers like “sale”, “discount”, “coupon”, “buy” or “buy online”.

Going back to our example, the searchers’ keywords can be “ridgeback coupons”, “ridgeback commuter bikes deals”, “ridgeback free shipping”, “ridgeback bike size guide”, “ridgeback commuter bikes comparison”, and so on.

The user intent is mostly transactional, with some commercial intent. Some navigational queries may occur when consumers check the manufacturers’ websites directly.

Tactics for this stage

  • Ensure your website ranks for branded search queries such as “{brandname}reviews”.
  • Your website will claim the first positions if you have pages that target reviews-related keywords and if you build just a couple of good backlinks from external sites to those pages.
  • A dedicated template page for “{brandname}reviews” will allow you to publish all the reviews for any product.

Figure 60 – This online retailer has a Reviews and News template for each brand.

Here’s how the website above ranks for “ridgeback reviews”. They got the #1 and #2 positions:

Figure 61 – If you own the brand, your site should easily come at the top, even without many backlinks.

Other tactics that you may consider are:

  • Distribute coupons to build links and brand awareness.
  • Write how-to content, user guides, and product comparison pages.
  • Optimize your brand pages and product descriptions to include reassurances, shipping estimates, refund policies, etc. Think in terms of optimizing your content for conversions rather than SEO.
  • Have a Promotions/Coupons/Reviews page targeting your brand terms.

Figure 62 – SERP for results for “Macy’s coupons”.

In the image above, you can see the results for “Macy’s coupons”. This is a great keyword to rank for, and Macy’s is ranked #2 for its brand name plus “coupons”. By creating this Coupon page on their website, they are taking away traffic from coupon websites.

  • If you accept coupon codes at checkout, make no mistake; consumers will leave the process to find your coupons. Instead of allowing users to leave the checkout to find current promotions outside your website, use a pop-up window or open a page in a new tab to list your current promotions and coupon codes.
  • Create interactive tools for finding, comparing, or visualizing products (e.g., virtual eyewear, try before you buy tools, see the painting in your room, etc.).

The purchase stage

This is the stage at which consumers know exactly what they want to buy or at least the brand they want to buy from. The queries contain specific product names and the exact model number or version of the product (e.g., Ridgeback Meteor 14). The keywords are the most focused at this stage. These are probably easier keywords to classify because they often contain the product name or the brand name. For ecommerce websites, the landing pages most associated with these queries are the product detail pages.

At this stage, the user intent is mostly transactional, with some navigational intent (for example, typing “Amazon” in a search engine to buy a book or purchasing directly from the manufacturer’s website).

Tactics for this stage

  • Engage appropriate influencers for product reviews and send qualified traffic to your website.
  • Develop backlinks to product detail pages.
  • Optimize product detail pages to include detailed product specs, persuasive descriptions, great images, questions and answers, etc.
  • Offer coupons.

The purchasing stage is the last in our buying funnel model. Some marketers and sales professionals have gone into greater depth and broken it down into even more detailed steps. However, if you start breaking down the funnel into four stages and begin developing content based on these stages, you will see traffic and sales increase nicely over time.

Keep in mind that a purchasing decision is never going to be linear. Prospective customers will start their journey in the middle or at the end of the funnel. Regardless of where the journey begins, you can capture consumers at any stage if your content is well-planned.

Knowing about the buying funnel stages is important for understanding another keyword research concept, the user intent.

The relationship between these two concepts is pretty tight. Usually, a consumer in the Awareness stage will use informational search queries. When the searcher is in the Purchasing stage, it mainly uses transactional intent keywords with strong commercial intent.

The user intent

Users try to accomplish something when they go to search engines and type queries. That something can be:

  • Finding a business that can be located either online or offline.
  • Getting more information about a product or a service.
  • Purchasing an item.

Searchers have a goal in mind, and that goal is called the user intent. Search engine users type in phrases representing their intents, and Google tries to match those intents with the most relevant results. If you understand this concept, you understand the importance of mapping keywords-to-intents and developing content accordingly.

Figure 63 – Three types of user intent keywords.

The specialty literature[8] breaks down the user intent into three categories:

  • Navigational – when searchers use a search engine to navigate to a specific website.
  • Informational – when searchers want to find content and info about a specific topic.
  • Transactional – when searchers want to engage in an activity, such as buying a product online, downloading or playing a game, seeing pictures, viewing a video, etc. Transactional intent does not necessarily involve a purchase.

Google’s guidelines for quality raters (the SERPs human evaluators[9] ) who assist with quality control of the SERPs refer to the same categories as Navigation, Informational, and Action.

When discussing user intent types, it is worth mentioning commercial intent. Commercial intent is an independent dimension that can apply to all three types of user intent, with transactional queries probably carrying a higher commercial intent than the other two. A Microsoft Research study found that 38% of the queries have commercial intent, and the rest are non-commercial.[10]

Figure 64 – Navigational and informational keywords can have commercial intent, too.

For example, when a consumer wants to buy a car, he will perform the research online, but he will seal the deal in a dealership. His queries, whether informational, transactional, or navigational, will all have some commercial intent because his final goal is to purchase a car.

Mapping keywords to intent is not an easy task. Even search engines cannot accurately classify user intent in general, let alone commercial intent. So, map keywords to user intent as best you can. As long as you start categorizing based on intent, you will begin generating ideas for content that matches the intent and is relevant to users. This is the best SEO approach to stand the test of continuous algorithm updates.

Below are some guidelines for classifying user intent, but remember that many keywords can be placed into multiple intent buckets.

These are queries containing:

  • Companies, brands, organizations, or people’s names.
  • Parts or full domain names.
  • The words “website” or “website”.

Navigational queries are the easiest to spot during keyword research. For this query type, ensure you appear at the top of the SERPs for your brand and domain name search queries. If you are not showing up at the top for such queries, you might have a more significant problem than mapping user intent. You might have a site-wide penalty.

Figure 65 – Best Buy pays for its brand name to appear on Google Ads. That is because they deemed the branded keywords very valuable.

If you sell someone else’s brands, it is not a good idea to put efforts into ranking for keywords made of brand names only because this means competing directly with the brand owners and their social media profiles. Overtaking them in rankings is not possible—unless the brand sucks in terms of SEO—and even then, it is going to require significant effort.

If you own the brand or are a manufacturer that sells your products, ensure that your website shows up for possible keywords containing your brand name and product names. For example, if you manufacture and sell computer RAM, your site should rank at the top for brand queries, including the products you sell (e.g., “Kingston 1Gb RAM” and “Kingston 1 Gb RAM”).

Informational intent queries, or the “know” queries

These are queries containing:

  • Question words (e.g., ways to, how to, what is, etc.).
  • Informational terms (e.g., list, top, playlist, etc.).
  • Anti-commercial queries (e.g., DIY, do-it-yourself, plans, tutorial, guide, etc.).
  • Words like instructions, information, specs.
  • Words like help, resources, FAQ.
  • A category or subcategories (e.g., digital cameras, raincoats, etc.).

If you have difficulty classifying keywords based on intent, one trick is to find the navigational and transactional intent queries first, then assume that the rest are informational.

If you want to learn how Google teaches its search quality raters to classify the search queries, I recommend reading Google’s “Search Quality Rating Guidelines”, especially Section Two.

Informational intent is the type of intent ecommerce websites should start shifting their attention to because informational keywords provide the chance to get in front of the target market in the early stages of the buying funnel. The earlier your audience is exposed to your brand, the higher the chances of closing a sale.

The content that addresses informational queries encompasses all media types, such as text, video, audio, etc. It includes product descriptions, technical specs, expert reviews, infographics, instruct-o-graphics, blog posts, how-to guides, etc.

When creating content for this type of intent, your goal is not to sell your products but to position yourself as the authoritative source in your space. You need to become a publisher of reliable and useful content. Informational queries are perfect because they represent an excellent opportunity to increase brand awareness and show expertise.

The fact that 80% of search queries are informational[11] represents a massive opportunity for those who plan for long-term gains. These queries can be very generic, for example, head terms such as category and subcategory names (e.g., “cars” or “insurance brokers”), but long-tail keywords as well. For example, search queries like “What is the most fuel-efficient car on the market?” or “Life insurance brokers in New Westminster, BC” are informational. Note that either of these two example queries could also have a transactional intent.

To cover as many informational queries as possible, you must create educational content for consumers who are not yet ready to buy or do not even know what they need. Your goal is to provide searchers with content that answers their questions and fills their need for information. Also, your content must assist in nudging searchers further down the buying funnel.

Informational intent queries appear at the Awareness, Research, and Decision stages. You must gradually guide consumers towards more transactional content, eventually leading to conversions. After all, a macro-conversion (e.g., a web purchase) happens only at the end of several micro-conversions, such as reading an article about a problem, finding the right product, adding to the shopping cart, clicking on proceed to checkout, etc.

One way to check whether there is a disconnect between user intent and the content on your website is by looking at the ecommerce transactions (and conversion rates) for your keywords (if the keywords data is available):

Figure 66 – This Google Analytics report became almost useless after Google stripped the search query data from the referring URLs (now showing as “not provided”).

You can look at which pages or keywords perform poorly using Google Analytics. In this example, getting 3.3k organic visits from a single keyword and ending with just one conversion indicates something is wrong. It may be because the searchers land on an improper landing page or because the landing page attracts the wrong keywords. It may also be related to conversion frictions, such as your pricing being higher than competitors.

Another way of finding this disconnect is by analyzing the keywords’ bounce rate (whenever you can get this kind of information):

Figure 67 – Whenever you can identify a keyword that drove traffic to the website, look at its bounce rate.

A high SERP bounce rate is usually bad because it shows that searchers landed on your website and didn’t find what they expected. However, blog pages typically have a high bounce rate since visitors might find the answer they want in the article and then leave.

As you start looking at keywords through the user intent prism instead of just words and numbers and try to solve high bounce rates, low conversion rates, and low transaction numbers, you will learn more about your visitors. This will help with organic traffic and everything marketing and sales.

When analyzing the performance of the informational queries, remember that such keywords will most likely not convert on the first visit.

The transactional intent or the “do” queries

These are queries containing:

  • Calls to action (subscribe, purchase, pay, play, send, download, buy, listen, view, watch, find, get, compare, shop, search, sell, etc.).
  • Entertainment terms (pictures, movies, games, and so on).
  • Promotional terms (coupons, deals, discounts, for sale, quotes).
  • Complete product names.
  • Comparison terms (where to buy, prices, pricing, compare prices).
  • Terms related to shipping (next day shipping, same day shipping, and free shipping).

However, not all transactional queries contain verbs. For example, the “Dell Vostro 1700” search query can be transactional and informational because the user wants to read more or buy it. Also, transactional queries do not necessarily have to involve money or purchases. They reflect only the desire to perform some action on the Internet.

Transactional queries with commercial intent occur more frequently in the decision and purchasing stages. Such keywords should land visitors on category and product detail pages or landing pages built to funnel visitors to a page where a commercial transaction occurs (e.g., a product comparison tool or a finder tool).

Transactional queries are most likely to generate the highest return on investment (ROI) for pay-per-click campaigns, which is why their cost-per-click can be high. However, the ROI would be even better if you had previously “touched” the pay-per-click searcher with an organic result. Upon landing on your website from the PPC ad, your brand might be recognized, which might positively affect conversions.

A possible way to connect user intent with search queries is using surveys sourced from your organic traffic. You can implement a modal window or a pop-up that tracks the search query used by visitors. The downside is that search engines no longer pass search query data in the URL string so that you will get only a fraction of the queries.

However, when you can identify the keyword, trigger the modal window and ask a simple question, such as “What is the goal of your visit to our website today?”. Provide two possible options:

  1. I am shopping for something to buy now or soon.
  2. I am looking for more information about some products/services.

Mapping intent to content

Your content strategy should be created by understanding where in the buying funnel the user is when he types in a search query, mapping his user intent, and bucketing the searcher into the right persona he belongs to.

But why is user intent so important? It is because an algorithm matches user intent with search queries: Hummingbird.

One of the metrics used by search engines to measure the match between the user intent, the search query, and the perfect result is SERP user engagement.

However, the ultimate metric that Google uses to quantify if the content on a page matches the user intent behind a search query. This metric is called The Long Click.

The following is a quote from the book “In the Plex: How Google Thinks, Works, and Shapes Our Lives” and describes the long click:

“On the most basic level, Google could see how satisfied users were. To paraphrase Tolstoy, happy users were all the same. The best sign of their happiness was the “Long Click” — This occurred when someone went to a search result, ideally the top one, and did not return. That meant Google has successfully fulfilled the query”.

Let’s start the keyword mapping process.

“Tired Jamie” is a persona developed for the scenario in which a buyer wants to purchase a mattress.

Scenario: Jamie cannot sleep at night and wants to improve her sleep. She finds that old mattresses can cause poor sleep and decides it’s time to buy a new one. She starts looking for information on choosing a mattress that can provide the best night’s sleep. She discovered a useful mattress finder tool that recommended foam mattresses based on her input. Next, she starts researching which brands sell foam mattresses and which have the best reviews. She found Tempur-Pedic®, which seems to be a trusted brand by many people, so she investigated their various types of mattresses. Finally, she knows what she wants and is actively looking for that product.

First, do your best to map her keyword journey by sorting keywords from top to bottom based on the buying funnel.

Figure 68 – This is the beginning of the keyword mapping process.

In this example, Jamie starts with a broad search, “tossing all night,” then refines it to “how to improve my sleep.” After discovering that mattresses can cause poor sleep, she refines her search to “how to choose a mattress”. Once she found the type of mattress that seemed to solve her problem, she started investigating “foam mattress brands”.

Once she finds a brand she trusts, she will look for their products by searching for “tempur pedic mattresses”; this search query contains the brand and the category of products. Finally, she looks for the specific product she intends to buy: “tempur pedic cloud luxe breeze”.

In the second column, tag each keyword by the most appropriate theme or silo it belongs to. For example, all keywords containing the word “mattress” will be bucketed in the Mattresses silo. “Tossing all night” and “How to improve your sleep” do not belong to a specific category of products so that you can assign them to a generic “resources” silo.

Next, map the keywords to the user intent while remembering that a query can have multiple intents. In our example, the first four keywords are informational, and the last two are transactional.

Then, add the type of content that fits the intent and the search query (e.g., for the search query “how to choose a mattress,” you can build a mattress finder tool).

Now, you need to add more details.

Figure 69 – We will keep adding more data to this table.

The URL represents the page you want to rank in the SERPs. This is the page you deem the most appropriate to rank with. In our example, for the keyword “tempur pedic cloud luxe breeze,” you will want to rank with this product page:

/tempur-pedic-cloud-luxe-breeze.html

In the Anchor Text(s) column, you will list the internal anchor text used to link to the targeted URLs (this can also be used as anchor text for your backlinks).

Note: failing to establish contact and presence with consumers who perform informational queries is one cause of single-digit conversion rates. Too often, ecommerce websites try to sell too early.

If you develop content for all the buying stages and intents, you can land prospective customers on your website at the research stage. Then, gradually nudge them to the purchasing stage without having them exit your website to find answers from competitors. If one of your competitors becomes the trusted source of advice for that potential customer, you have lost the sale.

So, ensure you optimize the right pages for the right queries. You want to rank a page with informational or educational content if the query is informational. Likewise, if the query is transactional, you need to optimize and rank with pages that have transactional intent.

Prioritization

Keyword prioritization is difficult because:

  • You need to consider many various metrics.
  • The SERP ranking factors are not publicly available.
  • Several metrics, such as competitiveness, come from third-party sources (not directly from the search engines).

Therefore, any keyword evaluation model based on ranking factors and competitiveness metrics is subjective. Prioritization methodologies are usually based on keyword difficulty, search volumes, business goals, profits, margins, conversion rates, or a combination of such metrics.

One lesser-known keyword prioritization method is based on forecasted rankings’ revenue opportunity. This evaluation model determines a monetary value for the top 10 rankings using the average SERP CTRs.

Note that this model is meant only as a tool to help you identify the lowest-hanging opportunities.

Figure 70 – We’re adding search and business metrics to the process.

In this table, you can get the Search Volume data with Keyword Planner, the Current Ranking data with the ranking tracking tool of your choice, and the Organic Visits data using your web analytics tool. The Revenue data is also collected from your web analytics tool. You will generate the Per Visit Value by dividing Revenue by Organic Visits.

I excluded metrics such as the conversion rate or the number of conversions on purpose. That is because ecommerce websites have multiple micro-conversions and macro-conversions (e.g., a web sale, a newsletter subscription, reaching a critical page, submitting a form, etc.), and this evaluation method is based solely on revenue, not conversion rates.

If you want to dig into details and evaluate based on each type of conversion (for example, prioritize keywords that generate more email subscriptions), then use only the newsletter revenue data for each keyword and prioritize accordingly.

The columns “Rev. if ranked 1…10” represent the revenue opportunity for various positions if you rank organically at those positions. Looking at the “Rev. if ranked #1” column, you can see that although the keyword “tempur pedic cloud luxe breeze” is a transactional keyword and has the highest Per Visit Value ($25), it is not the keyword with the highest potential to increase revenue. That keyword would be “how to choose a mattress”.

Figure 71 – SERP CTRs from Optify.

For this forecasting method, I used the organic SERP CTRs based on research done by Optify.[12]

The next step is adding keyword competitiveness data. There are a few different methods for assessing the competitiveness of a keyword. The easier ones are:

  • The average Domain Authority (DA) or Page Authority (PA) of the top 10 ranking pages and root domains. Note that the table below includes the average PageRank, but this data is not publicly available anymore.
  • MOZ generates the keyword difficulty score. The CI index is from serpIQ (now defunct).

Figure 72 – We’re adding competitiveness data.

Now that you have some quantitative information, you can slice and dice the keyword data any way you like. I suggest analyzing data in sets or themes only. If you mix keywords related to mattresses with dresser keywords, your data will be skewed.

Also, it is important to balance the forecasted revenue and the costs associated with obtaining the necessary rankings to achieve that revenue. Remember to produce content, promote it through various marketing channels, and build backlinks. All these actions have a cost.

Figure 73 – We’re adding costs related to producing and promoting content.

Content Creation Cost estimates how much it will cost to create the content necessary to promote the keyword. Each keyword might have a different cost depending on the type of content you need to create; creating an article is less expensive than creating a video, which is less costly than creating an interactive tool or a mobile app. The Cost per Link estimates how much it will cost to build one link to that content.

This is the Costs formula:

Costs=content creation cost + (cost per link * (average DA / 10)*2)

The number 2 in this formula is a cost coefficient tied to your domain authority. The lower the domain authority, the higher the coefficient. You can use the following brackets as guidelines for adjusting the coefficient based on your DA:

  • DA 0–20, coefficient=5
  • DA 21–40, coefficient=4
  • DA 41–60, coefficient=3
  • DA 61–80, coefficient=2
  • DA 81–100, coefficient=1

The Costs formula indicates that the lower your DA, the more links you need to build to achieve first-page rankings. For example, for the keyword “tempur pedic mattresses”, if your website DA is 65, the coefficient is 2, and the Costs formula is:

Costs = $ 250 + ($200 * (45/10)*2)=$250 + ($200 *9), where nine means you will need to build nine good-quality links.

You can download the Excel file containing the example formulas here.

Returning to user intent, remember that search engines aim to provide straight answers for search queries and the best possible results for keyword searches. If search engines fail at this, they will lose users, market share, and advertising revenue. It is, therefore, crucial for search engines to identify user intent as best as they can. Remember, Google changed its algorithm to focus on this with the release of Hummingbird. Microsoft used to have a publicly available commercial intent detector tool, but unfortunately, it has been discontinued. So, you will have to classify the results manually.

Whenever you doubt how search engines map keywords to intent, use Google’s help to assess what type of pages it returns for a specific keyword. First, log out of all your Google accounts. Then, clean out the browser cookies, open an incognito session, and type in the keyword you want to research.

For example, look at the SERP for the “digital camera” keyword. (see screenshot below). For this keyword, seven listings are informational or educational resources (non-commercial intent such as reviews, news, images, tips, wiki, etc.), and four are online retailers (commercial intent). Keep in mind that I counted image results as just one single result. If you sell digital cameras and want to rank for this keyword, you must create great educational resources on your website and promote them heavily on both your and external websites.

Given the number of informational results for “digital camera”, it seems Google does not assign a strong commercial intent to this keyword. Then why do ecommerce websites try to rank their Digital Cameras category URL rather than a Digital Camera page dedicated to educational content and tools? Wouldn’t a category page create a disconnect between the user intent and the content on that page?

Figure 74 – SERP for “digital camera”.

Imagine you walk into a store to get information about which digital camera best suits your needs, only to encounter a pushy salesperson who tries to sell you items he wants you to buy rather than what you think you need. You will probably thank him nicely and leave without buying. The same applies to online experiences; if searchers land on a page that does not fit their intent, they will bounce.

Creating content based on keyword research must address the possible buyers of your products and those who will link to your content. That is because most people who buy from you will not link to a product or category page. Customers may share the purchase socially, but back-linking will happen only from people who believe the content they link to is valuable. Buyers think about the value they get by buying the product from you, but those who link to you think about the value they offer to their audience.

Product attributes and keyword variations

Online retailers often sell products with similar attributes and would like to rank for many keywords and product variants. For example, you sell a red, blue, and green sweater in three different sizes (extra-large, large, and small). This matrix will generate nine product variants: small red sweater, large red sweater, extra-large red sweater, small blue sweater, large blue sweater, extra-large blue sweater, small green sweater, large green sweater, and extra-large green sweater.

In the section on product detail pages, we will discuss how you should approach product variants. However, for now, let’s say that creating unique product descriptions for each variant may not be the best idea unless you have a large budget for content creation. Instead, handle product variants in the interface without reloading the entire content (with URL changes). To achieve this, you can use dropdowns to allow users to pick a color and AJAX to load the content specific to that color variant.

If you already have unique URLs for product variants, choose a canonical URL and point all product variant URLs to it (be careful what you choose as the canonical version). However, if the URLs are clean (i.e., they do not have too many URL parameters), do not change the URL structure without consulting an SEO expert.

Figure 75 – A simple decision chart inspired by MOZ.[13]

Keyword strategies

This section will discuss a few less-talked-about keyword strategies for e-commerce websites.

Target low-hanging fruits

When you run an ecommerce website, the number of keywords you want to rank for is enormous, so it is not economically feasible to target all of them with link-building campaigns. You can rank organically for many long-tail keywords by supporting them with content. Other keywords (usually the more competitive terms such as category and subcategory names) will only rank if you support them with content-rich website sections and links from external websites.

An often-overlooked keyword strategy is to focus on keywords ranked on page 2, especially those ranking between positions 11 and 15. Moving a keyword from the second to the first page is usually less competitive than moving the same keyword four spots up, from 5 to 1. In the same way, moving from position 21 to 17 (also four positions up) will not generate a substantial increase in visits.

Let’s illustrate this concept with a keyword with 1.2 million monthly searches, “wedding dresses”. Moving this keyword from position 11, where it gets less than 2.6% of the clicks, which is about 3,100 visits, to position 6, where it gets 4.1% of the clicks, which is about 4,900 visits, represents a 158% improvement in traffic. Moving the same keyword from position 21 to 16 will generate a minimal rise in visits.

Figure 76 – Keywords ranking at the top of page 2 can be a good target for link development

The idea is that by building links to keywords ranking on the second page, you gradually increase your website’s authority and, at the same time, generate more organic traffic. These links will eventually support the link-building campaign for more competitive terms.

Of course, you should not focus solely on keywords ranking on the second page. A thorough analysis will identify keywords that rank on the first page and do not have much competition—it also makes sense to target those.

Target holidays and retail days search queries
Holidays such as Christmas, Hanukkah, Thanksgiving, Easter, and major retail events such as Back to School, Halloween, Cyber Monday, Black Friday, and Boxing Day represent significant traffic and revenue opportunities for all e-commerce websites and online retailers. Shoppers are more open to spending during holidays. However, their search patterns change around these special shopping days.

The ecommerce shopping days calendar created by Shopify[14] shows no month without a major shopping event. Promotions change rapidly, and shoppers will shift their search queries very quickly. Smart ecommerce websites have to adapt to and capitalize on such shifts. However, many websites do not have the agile SEO abilities to take advantage of shopping day opportunities.

Here are some common mistakes made by online retailers regarding targeting shopping events:

  • Not updating page titles, descriptions, and headings to include event-related modifiers.
  • Not updating page titles, descriptions, and headings to include event-related modifiers until just a few days before that event. This is too late from a business point of view and for SEO. Google Insights research[15] suggests that Black Friday searches can come as early as July.

Figure 77 – 30% of shoppers plan their Christmas shopping list before Halloween.

  • Creating year-specific pages (e.g., Christmas 2018) and removing them without proper redirects once the holiday/event ended.
  • Not planning a “flash” link building campaign to target event-specific modifiers. Regarding link building, a flash campaign means two to three months in advance.
  • Not targeting last-minute buyers by adding “free” or “next-day shipping” in the page titles.

Figure 78 – 55% of consumers expect free shipping.[16]

Additionally, very few ecommerce websites will create content (e.g., guides, ideas, how-to’s) specifically targeting such retail dates. That is a shame since this type of content can capture potential customers during their research stage, when they use informational searches queries, like “Halloween costumes ideas”, “Christmas gift guides” or “Easter egg decorating pictures”.

Use keyword modifiers to update titles, descriptions, and headings
The way consumers search online before and during shopping events differs from how they do so the rest of the year. They add keyword modifiers to their usual search queries to better define their intent. Event modifiers are words like “Christmas”, “Cyber Monday” or “Boxing Day”, but also “same day shipping”, “next day shipping” and even “gifts”.

Look at the spikes in search volume associated with the “next day shipping” keyword modifier. The peaks reach the maximum a few days before Christmas. You should be fast enough to capitalize on this search pattern change.

Figure 79 – Shipping-related queries increase significantly around Christmas.

Figure 80 – Adding “same-day shipping” or “next-day shipping” to your titles by December 15 may prove wise.

Let’s say you want to capitalize on searches that contain the keyword “Christmas”. Add the word “Christmas” in the title of the category or product detail pages immediately after Cyber Monday is over. You can also consider altering meta descriptions and page content as well. Be sure to check the rankings associated with these pages a couple of days after you made the updates (and regularly after that), to see whether there is a traffic drop. You can expect some fluctuations, but as you get closer to Christmas, you should see an increase in rankings and traffic.

If there is a drop, revert to the usual titles. If there is an increase, change all the titles for the category, subcategory, and product detail pages. As you get closer to Christmas (e.g., December 15), change the title to “Free same-day Christmas shipping” since “free shipping” tops the list of the strongest incentives for visitors to buy goods online.

Get any page crawled and indexed by Google in less than one minute
Use the Fetch as Google feature in Google Search Console to achieve this. In the new GSC, use the Test Live URL functionality.

Figure 81 – Once you hit the Fetch button, Googlebot will crawl the submitted URL. If the page passes Google’s filters, it will be indexed in minutes.

Figure 82 – The number of fetch requests in Google Search Console is limited, so use your quota wisely.

Once Christmas ends, change the titles to target the next Holiday, e.g., Boxing Day. If there is a gap of more than three to four weeks between shopping events, you can default to the usual titles.

You may want to use an automated system that allows event-specific titles, descriptions, and headings to be updated on specific dates. If that is not possible, then at least set up calendar reminders a month before the less important shopping events and two months before the most important ones. You can refer to this article[17] for consumer trend data and the importance of each consumer holiday.

Create holiday-specific landing pages.
To drive targeted traffic, marketers create holiday or promotion-specific landing pages with PPC, email, and catalogs. During the year, they will create pages for “Christmas 2018 Promotion”, “Father’s Day Specials” or “Valentine’s Day Two-for-One Deals”. You have probably noticed this implemented by big brands or small but smart competitors.

By creating these landing pages, which are visually themed per the event they target, marketers make ecommerce websites more attractive to visitors. These pages can get natural links from deals or coupon websites if your brand is recognizable or you push the pages with an outreach campaign.

Usually, when e-commerce websites use specific event or holiday landing pages, they publish them on URLs such as mysite.com/Boxing-Day-Sale. However, improper redirect handling, such as no 301 redirects or 301 redirects to the wrong pages, may lead to PageRank loss once the event ends. In such mishandling, the pages are removed from the website.

Here are a couple of tips if you use separate URLs for holiday or shopping events:

  • Do not include years or any other time or date indicators in URLs. Including time indicators in page titles, descriptions, headings, and main content is okay.
  • When the event is over, redirect the event pages to the most appropriate website sections or keep the URLs alive (but with changed content).
  • You can “revive” the promotion-specific URLs a few weeks before each event in the following years.

A mixed approach
Whenever possible, I like to implement another tactic: updating titles, descriptions, and headings while customizing the look and feel of the existing landing pages. Instead of having separate URLs for each event, your current landing pages (for example, your category pages) will become the event landing page.

Choose the most important categories on your website or those promoted during a specific consumer holiday and customize their look and feel to match the event. Customization can be as simple as displaying a banner at the top of the main content area or adding a background image for the entire page, or it can be as complex as creating an entirely new event-themed layout.

You will not publish this new layout under a different URL. Instead, this themed look, feel, and messaging will be released on the regular category page URLs. For example, your Christmas 2018 promotion includes a 25% discount on all Cleansing products. Rather than creating the URL mysite.com/Christmas-Deals for this holiday, you will use its usual URL, mysite.com/Cleansing/. However, this page will be themed with a Christmas look and feel.

The main benefit of customizing the existing category pages for shopping events is that you can build backlinks to category pages more easily. Another benefit is that there will be no future redirect headaches. Also, if other websites want to link to your promotions, they will link to your category pages.

Once the holiday ends, return to the usual, non-themed layout. You will also have to update titles, descriptions, and page copy (a bit).

Tip: If your website gets image-rich snippets, you can “theme” the image thumbnails with an event or holiday-specific icon. Once the holiday/event is over, revert to the usual images. For instance, if you sell cameras, instead of this video thumbnail:

Figure 83 – Video listing in search results.

Use a themed image:

Figure 84 – Personalizing the video thumbnails can lead to a better CTR.

Optimize for “gift card” related keywords

Last-minute shoppers often buy e-gift cards instead of real products to avoid shipping delays. If you are still debating using gift cards, consider the following:

  • According to Giftango Corp, 26.7% of the gift cards sold during December 2011 were sold between December 21 and 24.[18]
  • 57.3% of shoppers planned to buy a gift card in 2011.[19]
  • Gift cards were the most requested gift in 2012, with 59.8% of US shoppers wanting one.[20]
  • E-gift cards reach their recipients instantly (no delays, no shipping, and no hassles).

It makes sense to offer both e-gift cards (perfect for last-minute shoppers) and gift cards (great for those who do not know what to buy as gifts).

Target long-tail keywords
For ecommerce websites (especially those new on the market), it is more viable to start by targeting long-tail search queries and gradually progress towards more competitive head terms. Usually, carefully chosen torso and long-tail search queries tend to generate more qualified traffic and have less competition. However, keywords containing brand names may prove as competitive as head terms.

Targeting search queries that assist with conversion is a good tactic. Often, such search queries require content (interactive tools, comprehensive guides, etc.), but think of this content as a long-term investment. For example, targeting the search query “how to choose a digital camera” may require creating a camera finder tool. If you target “how to choose shaving cream,” you must create an extensive (eventually interactive) and visually appealing resource specifically for that.

Here are just a few benefits of targeting long-tail keywords:

  • You will achieve organic search results fast.
  • It helps gather insights about your customers.
  • It assists with improved paid search results (through better Quality Scores).

Figure 85 – This is the SERP for the query “how to choose shaving cream”. None of the top 10 results has a product finder or a product wizard. If you are in this niche, that is your opportunity.

In addition to focusing on long-tail search queries, you may need to avoid targeting head terms with very vague user intent. For example, let’s say you sell greeting cards. Would ranking for a keyword like “greeting” or “cards” be useful? No, because you cannot identify the user intent behind these keywords. You will invest a lot to brand your business for those terms, and you will get a ton of traffic if you rank at the top, but generic terms generate very few conversions at a very high cost per conversion. Instead, you can start targeting keywords like “40th birthday greeting cards for dads”, perhaps on a blog post or, if it is a popular search query, with a content-rich subcategory page.

As you can see, keyword research is far from simple or fast. It is a process that cannot be fully automated, and human review is irreplaceable, especially when bucketing keywords for relevance to your business. After going through this section, you hopefully understood that performing keyword research without considering user intent is bad.

During the next sections, we will find that keywords are part of almost every on-page SEO factor, from page titles to URLs, internal anchor text, and product copy. However, for search engines to find and analyze keywords, they must first find and reach the pages where those keywords are featured. Since ecommerce websites are a challenging crawling task for search engines, you must optimize how search bots discover relevant URLs. This process is called crawl optimization, and it is the subject of the next section of the guide.

References: